Service recommendation is an active research area in cloud computing since mapping of right services as per the desired requirements of user is a challenging task. With the rapid development of cloud services, the recommendation techniques are facing the problem in predicting the better QoS values because of neglecting the contextual information of cloud services. The paper proposes a service context-aware-based cloud broker that extracts the service details by considering the contextual information of cloud services and computes service similarities on the basis of QoS values. Further, it tackles the cold start problem by adopting the matrix factorization principle and predicts better QoS values for newly arrived services. To validate our approach, we have conducted experimental works on benchmark datasets and the result shows that the proposed approach outperforms better results than the model-based approaches. In particular, our proposed system produces improved response time for the dataset of sparse nature.